Slopsquatting
Named by 18 kept items in this workspace.
Search every workspace for Slopsquatting. It matches the name as written, so it can find other things that share it.
The items that name it 18
Attackers exploit AI coding tools that hallucinate dependency names across ecosystems (e.g., assuming a Python package like 'requests' exists in npm, or Axios-equivalents named 'requests'/'fetch'), making typosquat-style attacks more effective.
An attacker registers a made-up (hallucinated) package name so the next agent requesting it installs real malware
Hallucinated or outdated dependencies where a model suggests packages that are unsafe or do not exist
Attackers exploiting hallucinated (non-existent) package names suggested by AI coding assistants
A software supply chain attack that turns AI hallucinations into malware delivery; an attacker registers a package name that an AI coding tool hallucinates, and the next developer who follows the AI suggestion installs the attacker's package instead of a real one (a 'pull' attack, since the victim's
Registering plausible-but-hallucinated package names that AI tools suggest, to weaponize against developers or coding agents
Attack where adversaries register malicious packages under names AI models hallucinate; the page cites nearly 20% of AI-generated package recommendations referencing nonexistent packages
malicious package technique targeted for blocking at the developer-machine and organization level
WEL1DROPPER packages are forks of benign projects poisoned and published under names likely to be suggested by AI coding assistants, targeting agents rather than humans
Software supply chain attack where an attacker registers a nonexistent package name that an AI coding agent hallucinated, loads it with malicious code, and waits for a developer/agent to install it
Attackers register plausible-sounding package names hallucinated by AI models and load them with malware
Squatting attack built on package names that AI coding agents invent outright, which developers accept without verifying
Attack technique exploiting AI-hallucinated package names by registering malware under a name an LLM repeatedly invents, so AI coding tools recommend it to developers
Attackers register package names that AI coding agents are likely to hallucinate, so agents install a pre-prepared malicious package when they guess wrong.
Attackers guess package names that AI coding tools are likely to hallucinate and publish malware under those names, so that when an AI 'invents' a plausible but nonexistent package, it actually resolves to attacker-controlled malicious code
Attackers register package names that AI coding tools hallucinate; when a developer accepts an AI-suggested dependency that doesn't exist, they may install an attacker-claimed malicious package instead.
AI slopsquatting campaign (WEL1DROPPER)
All techniques in this workspace.